What’s New in Version 7.6
Couchbase is the modern database for enterprise applications. Couchbase Server 7.6 combines the strengths of relational databases with the flexibility, performance, and scale of Couchbase.
For information about platform support changes, deprecation notifications, notable improvements, and fixed and known issues, refer to the Release Notes.
New Features and Enhancements in 7.6.2
The following new features are provided in this release.
Platform Support
-
Couchbase Server 7.6.2 adds support for the following platforms:
-
Ubuntu 24.04
See Supported Platforms for a full list of supported platforms.
-
Backup
-
Users with the Read-Only Admin role can now read backup information from the following Backup Service REST API endpoints:
-
You can now set the number of threads each Backup Service node uses when backing up data. For example, if you find backups cause performance issues on your cluster, you can reduce the number of threads the Backup Service uses. Reducing the number of threads also reduces the number of concurrent client connections the Backup Service makes to retrieve data. See Thread Usage for more information.
Analytics Service
-
The Analytics Service REST API has two new endpoints that let you get information about active and completed requests. See Active Requests and Completed Requests in the Analytics Administration REST APIs page.
Cluster Manager
-
Version 7.6.2 adds Cluster Manager metrics to help you monitor failovers and rebalances:
-
cm_auto_failover_count
: the number of auto-failovers that occurred. -
cm_auto_failover_enabled
: whether auto-failover is enabled. -
cm_auto_failover_max_count
: the maximum number of auto-failovers allowed before Cluster Manager disables auto-failover. -
cm_failover_total
: The total number of non-graceful failovers that have occurred. -
cm_graceful_failover_total
: The total number of graceful failovers that have occurred. -
cm_is_balanced
: Whether the Cluster Manager is balanced. Only reported by orchestrator nodes and only reported every 30 seconds. -
cm_rebalance_in_progress
: Whether there is a rebalance occurring. Only reported by the orchestrator node. -
cm_rebalance_progress
: An estimate of the progress of the current rebalance. Only reported by the orchestrator. -
cm_rebalance_total
: The total number of rebalances that have occurred.
For more information, see Cluster Manager Metrics.
-
Query Service
-
In Couchbase Server version 7.6.2 and later, language constructs that may allow for code injection, speculative execution attacks, or side channel attacks have been removed from JavaScript user-defined functions in the Query service. For details, see JavaScript Functions for Query.
Index Service
-
From version 7.6.2, you can specify that index creation operates in deferred build mode by default. In deferred build mode, creating an index does not trigger the index build phase: you must trigger the index build before you can use the index. For details, see CREATE INDEX.
-
In Couchbase Server Versions 7.6.0 and 7.6.1, enabling file-based index rebalance prevented you from controlling which Index Service nodes contain an index. Version 7.6.2 removes this restriction. You can now use the
WITH <node>
clause of theCREATE INDEX
SQL++ statement when your cluster has file-based index rebalancing enabled. See Index Rebalance Methods for more information.You still cannot use the WITH <node>
clause with theALTER INDEX
statement when your cluster has file-based index rebalancing enabled.
Search Service
-
Version 7.6.2 adds multiple improvements to the Search Service, including Vector Search:
-
Vector Search now automatically adds a match_none query object to your Vector Search queries. This means you can run a pure Vector Search without adding a
query
object to your JSON Search request. -
The total dimension support for Vector Search has been increased from 2048 to 4096 array elements.
-
Vector Search now supports vectors encoded as efficiently compressed base64 strings.
Add a base64-encoded vector to your Search index with the new
vector_base64
field data type, then use thevector_base64
object in your Search request. -
Use the Couchbase Server Web Console to edit and generate a full curl command for a Search request, with a new built-in code editor. Use the curl command to run a Search query with the Search REST API.
-
The Search Service can now index and search for metadata stored in Extended Attributes (XATTRs) fields inside your documents.
Use the new XATTRs mapping type in a Search index, then add the prefix
_$xattrs.
to a field object in your Search request. You can also use the Search Functions from SQL++ to search for XATTRs data.You must use the META function to select the XATTRs field you want to search for an uncovered query, or any SQL++ query without a suitable Search index. Use the SEARCH_META function to select XATTRs fields when you do have a suitable Search index. -
Vector Search is now supported on MacOS.
-
Vector Search is up to 7 times more performant, due to improvements in using node resources.
-
Eventing Service
-
Version 7.6.2 adds the following improvements to the Eventing Service:
-
The new base64 Encode and Decode functions let you pack large arrays of floats as efficiently compressed strings when you use Eventing to generate vector embeddings.
-
The new Sub-Document LOOKUPIN and Sub-Document MUTATEIN operations let you search for or modify a specific field in a document without having to search and retrieve the entire document.
-
You can now fetch and modify a document’s Extended User Attributes (XATTRs) through the Sub-Document LOOKUPIN and Sub-Document MUTATEIN operations.
-
SDKs
-
Alongside Version 7.6.2, Couchbase announces the 1.0.0 GA release of the C++ SDK. This SDK has been long used as the core of our Node.js, PHP, Python, and Ruby SDKs, to handle communicating with the cluster over Couchbase’s binary protocols. and is now released as a full, standalone SDK for applications needing the speed of C++.
-
Version 7.6.2 adds the ability to use multi-document ACID transactions with binary documents, alongside the current handling of JSON documents. This feature is initially implemented in the C++ and Java SDKs.
Tools
-
From version 7.6.2, the Couchbase Server tools package includes the
cbq
shell, alongside the previously-provided tools such ascbimport
,cbexport
, andcbbackupmgr
. For details, see CLI Reference.
New Features and Enhancements in 7.6.0
The following new features are provided in this release.
Platform Support
-
Couchbase Server 7.6 adds support for the following platforms:
-
Alma Linux 9
-
Debian Linux 12 (Bookworm)
-
Rocky Linux 9
-
macOS 13 "Ventura"
-
macOS 14 "Sonoma"
See Supported Platforms for a full list of supported platforms.
-
-
In response to CVE-2023-5363 and CVE-2023-5678, OpenSSL upgraded to version 3.1.4.
This update changes the available ciphers for TLS connections. If you have not updated your client applications to use recent TLS libraries, you may experience an inability to connect and TLS handshake failures. Before upgrading, we recommend testing compatibility in a separate environment – especially if you are unsure that your platform TLS (OpenSSL, Java Secure Socket Extensions, .NET Security Provider, etc.) has compatible ciphers.
Cluster Manager
-
A required minimum can be established for the number of replicas configured for a bucket. See Setting a Replica-Minimum.
-
In each user-created or sample bucket, a
_system
scope is created and maintained by default. This scope contains collections used by Couchbase services, for service-specific data. See _system Scope and its Collections. -
A rank can be assigned to each bucket on the cluster, whereby each bucket’s handling by the rebalance process is appropriately prioritized. Assignment can be made by means of either the CLI or the REST API. This feature allows the cluster’s most mission-critical data to be rebalanced most quickly. See Creating and Editing Buckets.
-
You can now have Couchbase Server prune rotated audit logs after a period of time. You set how long Couchbase Server should keep audit logs by using the new
pruneAge
parameter for the/settings/audit
endpoint. The default value of 0 means that Couchbase Server does not prune audit logs. See Configure Auditing. -
You can add one or more arbiter nodes to a cluster. An arbiter node helps your cluster in two ways:
-
It provides fast failover which helps decrease the cluster’s latency when reacting to a failover.
-
It provides quorum arbitration that helps avoid contention issues if the nodes in the cluster become partitioned.
-
-
The
sampleBuckets/install
REST API method now returns a JSON object containing the list of tasks Couchbase Server started to load the buckets. In addition, the/pools/default/tasks
REST API endpoint now takes an optionaltaskId
parameter to view details about a sample bucket loading task. See Install Sample Buckets with the REST API for more information.
Backup and Restore
-
The Role-Based Access Control (RBAC) REST API has a new
backup
endpoint that lets you backup and restore user and user groups. See Backup and Restore Users and Groups. -
The
cbbackupmgr
command has a new--enable-users
flag that backs up user groups and users including roles and permissions. When you supply the new argument,cbbackupmgr
saves user passwords in the backup in a hashed format. When restoring a backup,cbbackupmgr
defaults to not overwriting existing users in the database with identically named users in the backup. You can change this default behavior using the new--overwrite-users
command-line argument. See cbbackupmgr config and cbbackupmgr restore for more about user backup. -
The
cbbackupmgr
encrypted backups feature is now GA for both cbbackupmgr CLI and the Backup Service. See Backup Encryption.
Cross Data Center Replication (XDCR)
-
Node-connectivity can be checked, prior to the creation of an XDCR reference. See Checking Connections.
-
Binary documents can optionally be included in, or excluded from XDCR replications. See Filtering Binary Documents.
Performance
-
You can now migrate buckets from one storage backend to another. This feature supports migrating buckets from Couchstore to Magma and from Magma to Couchstore. You can migrate buckets while the database continues running. To complete the migration you must trigger a swap rebalance or a graceful failover followed by a full recovery on each node that contains the bucket. See Migrate a Bucket’s Storage Backend.
Security and Authentication
-
Security settings now provide additional parameters, for the configuration of Couchbase-Server user-password hashing. See Configure On-the-Wire Security.
-
Credentials for Couchbase-Server internal users can now be rotated at any time, by means of the REST API. See Rotate Internal Credentials.
-
LDAP authentication now supports using regular expressions to map users to LDAP users and groups. You can supply multiple regular expressions that Couchbase attempts to match against the user name supplied during an authentication attempt. This feature gives you greater flexibility when authenticating users. For example, you can use a regular expression to map the domain name in an email address to an LDAP organization. See Advanced Query under User Authentication Enablement.
-
The Couchbase Server Web Console now supports using Structured Authentication Markup Language (SAML) for authentication. When you enable SAML authentication, a Sign In Using SSO button appears on the Web Console login screen. This button lets users who have already authenticated with the SAML identity provider (Okta, for example) to skip having to enter credentials. See SAML Authentication for more information.
-
Couchbase Server’s LDAP support now has a setting that turns on and off TLS middlebox compatibility. This setting controls low-level network communication options when Couchbase Server securely connects to an LDAP server through intermediate systems such as proxies and firewalls. See Advanced Settings on the Configure LDAP page for more information about this setting.
-
Couchbase Server now supports using Public-Key Cryptography Standard (PKCS) #12 format certificates for node certificates. This format lets you bundle the node’s private key, public key, and certificate chain into a single file. See PKCS #12 Certificates for Nodes for more information.
-
Couchbase Server now supports the X.509 Elliptic Curve Key cipher suites. Elliptic Curve Key ciphers are less resource-intensive than other cipher suites. They’re useful when communicating with resource-constrained devices such as IoT hardware. See Private Keys for more information.
-
Couchbase Server no longer supports TLS versions 1.0 and 1.1. When upgrading to version 7.6 or later, the upgrade process automatically sets
minTLSVersion
totlsv1.2
if it’s set totlsv1
ortlsv1.1
. Before you upgrade, be sure all the clients you use support TLS 1.2 or greater. See On-the-Wire Security for more information. -
To prevent LUCKY13 attacks, Couchbase Server 7.6 removes the following ciphers from the default cipher list:
-
TLS_ECDHE_ECDSA_WITH_AES_256_CBC_SHA
-
TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA
-
TLS_RSA_WITH_AES_256_CBC_SHA
-
TLS_RSA_WITH_AES_128_CBC_SHAa
-
-
You can now enable alerts for certificate expiration. When enabled, Couchbase Server alerts you when server, node, or XDCR certificates are within 30 days of expiration. You can change the alert period via the new
certExpirationDays
alert limit setting. Couchbase Server sends a second alert when certificates expire. See Certificate Expiration for more information. -
Couchbase Server now defaults to using the (Argon2id algorithm to hash passwords for new users. This hashing algorithm is more secure than the SHA1 algorithm used to hash passwords in earlier server versions.
-
If you upgrade a database to Couchbase Server 7.6 or later, it continues to use the older SHA1 hashing algorithm for existing user passwords. You can enable a new setting that has Couchbase Server migrate user passwords from SHA1 to Argon2id when a user authenticates. This setting works only if the entire cluster is running Couchbase Server version 7.6 or later. For more information, see Automatic Password Hash Migration.
Metrics
-
Couchbase Server has a new service discovery endpoint to help you configure the Prometheus event monitoring system. The old endpoint, named
/prometheus_sd_config.yaml
is now deprecated. The new endpoint is able to produce the same output as the old endpoint and has additional features. See Configure Prometheus to Collect Couchbase Metrics. -
Disk usage statistics now include transient files in progress, state files, and configuration files.
Index Service
-
You can choose to have the rebalance process move an index’s files between nodes instead of rebuilding them from scratch. This setting improves rebalance performance as moving the files is faster than rebuilding them. See Index Rebalance Methods.
Search Service
-
Couchbase Server 7.6 introduces Vector Search to enable AI integration, semantic search, and the RAG framework. A developer-friendly vector indexing engine exposes a vector database and search functionality. With Couchbase Vector Search, you can enable fast and highly accurate semantic search, ground LLM responses in relevant data to reduce hallucinations, and enhance or enable use cases like personalized searches in e-commerce and media & entertainment, product recommendations, fraud detection, and reverse image search. You can also enable full access to an AI ecosystem with a LangChain integration, the most popular open-source framework for LLM-driven applications.
A Vector Search database includes:
-
Standard Couchbase vertical/horizontal scaling
-
Indexing capable of efficient Insert/Update/Removal of Items (or documents)
-
Storage of raw Embedding Vectors in the Data Service in the documents themselves
-
Querying Vector Indexes (REST and UI via a JSON object/fragment, Couchbase SDKs, and SQL++)
-
SQL++/N1QL integration
-
Third-party framework integration: LangChain (later LlamaIndex + others)
-
Full support for Replicas Partitions and file-based Rebalance
-
Vector Search is currently only supported on Couchbase Server 7.6.0 deployments running on Linux platforms. MacOS and Windows platforms are not supported. |
For more information about vector search, see Use Vector Search for AI Applications
Data Service
-
Introduces KV Range Scan, used to retrieve all documents in a specified range directly from the Data service. Note that in this initial version, you will achieve better performance using a direct fetch or retrieval from a Query with an Index. See the SDK docs for more information.
-
Two changes in Couchbase Server 7.6 affect the
maxTTL
setting for collections:-
In earlier versions, you could only set a collection’s
maxTTL
setting when creating the collection. You can now change themaxTTL
setting on a collection after creation. -
You can now set a collection’s
maxTTL
to -1 to prevent a bucket’s non-zeromaxTTL
setting from causing documents in the collection to expire automatically. This new setting is useful if you want most of the documents in a bucket to automatically expire, but want to prevent the documents in one or more collections from expiring by default.
See Expiration for more information.
-
Query Service
-
SQL++ language additions:
-
OFFSET clause added to the DELETE statement. See DELETE.
-
GROUP AS clause added to the GROUP BY clause. See GROUP BY Clause.
-
FORMALIZE() function. See FORMALIZE().
-
Multi-byte aware string functions. See String Functions.
-
Support for sequences. See Sequence Operators.
-
EXPLAIN FUNCTION statement. See EXPLAIN FUNCTION.
-
-
cbq shell additions. See cbq:
-
The
-query_context
command line option. -
The
-advise
command line option.
-
-
The WITH clause adds support for recursive CTEs. See WITH RECURSIVE Clause.
-
The CREATE COLLECTION statement adds support for maxTTL. See CREATE COLLECTION.
-
The
/clusterInit
endpoint in the Nodes and Clusters REST API adds support for Query memory quotas. See Initializing a Cluster. -
Named and positional parameters can now be prefixed by
$
or@
in a query. See Named Parameters and Positional Parameters. -
The
system:indexes
catalog now enables you to find the number of replicas configured for each index. See Query Indexes. -
The Query Service adds cluster-level and node-level parameters to limit the size of explain plans in the cache. See queryPreparedLimit and prepared-limit.
-
The Query Service adds support for sequential scans, controlled by RBAC, which enables querying without an index. See Query without Indexes.
-
The node-level N1QL Feature Control parameter now accepts hexadecimal strings or decimal integers. See n1ql-feat-ctrl.
-
Queries can now read from replica vBuckets when active vBuckets are inaccessible. The Query service adds new cluster-level, node-level, and request-level parameters to configure this feature. See Query Settings.
-
The CREATE FUNCTION statement now enables users to create a SQL++ user-defined function and the corresponding external JavaScript code in a single operation, without having to create an external library. See SQL++ Managed User-Defined Functions.
-
When a query executes a user-defined function, profiling information is now available for any queries within the UDF. See Manage and Monitor Queries.
-
The Query service collects statistics for the cost-based optimizer automatically when an index is created or built. See Understand the Cost-Based Optimizer for Queries.
-
The ORDER BY and GROUP BY operations overspill to disk if they exceed the Query service memory quota.
Eventing Service
-
The optional parameter
{ "self_recursion": true }
can be used with the INSERT, UPSERT, and REPLACE advanced operations to prevent the suppression of recursive source bucket mutations. For more information, see Optional{ "self_recursion": true }
Parameter. -
The built-in
ANALYTICS()
function allows the Eventing Service to integrate directly with SQL++ Analytics. This integration simplifies Eventing code logic and lets Eventing benefit from the high availability and load balancing of SQL++ Analytics. For more information, seeANALYTICS()
Function Call. -
The advanced TOUCH operation allows you to modify the expiration time of a document without having to access that document first. For more information, see Advanced TOUCH Operation.
-
The Sub-Document MUTATEIN operation allows you to modify only parts of a document instead of the entire document. This Sub-Document operation is faster and more efficient than a full-document operation like REPLACE or UPSERT. For more information, see Sub-Document MUTATEIN Operation.
Analytics
-
Power BI Connector version 1.0 released. (Power BI Connector documentation)
You can download the installation package from the following location:
Binaries Binaries SHAs
Install & Upgrade
-
Due to an Erlang compatibility issue, you cannot directly upgrade to Couchbase Server 7.6 from version 6.5 through 7.0. To upgrade a database running one of these earlier versions to 7.6, first upgrade it to Couchbase Server 7.1 or 7.2. See Upgrade for more information.